AI companies are buying, cutting apart and scanning millions of physical books to train chatbots, according to a Washington Post report based on newly unsealed court filings. The documents reveal Anthropic's secret "Project Panama" and shed light on similar book-acquisition efforts by Meta, OpenAI and Google, as the AI industry's race for high-quality training data fuels an escalating copyright battle.

Anthropic muss 1,5 Milliarden Dollar zahlen – wegen illegaler Buchdownloads. Doch das KI-Unternehmen hat bereits eine neue Strategie, um kostengünstig an Text zu gelangen:…

Its pitch is simple: older books offer cleaner data. "The world's best AI training data is sitting on a shelf," the company says on its website. "Books...